File system performance indicators

File system performance indicators are measures used to judge how well a file system works:

1. Throughput (data rate): amount of data transferred per second (KB/s or MB/s), depending on block size and the allocation scheme.
2. Response time / latency: time from issuing a request (open, read, write) to its completion.
3. Disk access time: seek time + rotational latency + transfer time; reduced by good layout (contiguous blocks), scheduling and caching.
4. Cache (buffer) hit ratio: fraction of requests served from the block cache without disk access; higher means better performance.
5. Disk space utilization: percentage of disk space used by useful data; affected by block size, fragmentation and metadata overhead.
6. Fragmentation: internal (unused part of last block) and external/file fragmentation, which increases seeks.
7. Number of disk accesses per operation (including directory lookup and i-node reads).
8. Reliability and recoverability: consistency after crash, and scalability to large disks and files.
9. CPU overhead: CPU time consumed by file system operations.

Techniques to improve them: block caching, read-ahead, reducing arm motion (placing i-nodes in the middle of disk or in cylinder groups), defragmentation and a proper block size.